Development of a Clinical-Based Prediction Model for Pediatric Intussusception: A Prospective Observational Study.

Area of Science:

• Pediatric Emergency Medicine
• Clinical Prediction Modeling
• Gastrointestinal Emergencies

Background:

• Intussusception is a critical pediatric condition causing bowel obstruction and ischemia.
• Accurate and timely diagnosis of intussusception is vital for effective treatment.
• Ultrasound is the primary diagnostic tool for intussusception in children.

Purpose of the Study:

• To develop a clinical prediction model for intussusception.
• The model utilizes history, symptoms, and physical examination findings.
• The goal is to identify pediatric patients requiring ultrasound evaluation for intussusception.

Main Methods:

• Prospective observational study conducted in a pediatric emergency department.
• Included children under 6 years old with suspected intussusception.
• Logistic regression was used to identify predictors and develop the model, with diagnostic performance assessed using sensitivity, specificity, PPV, NPV, and AUROC.

Main Results:

• Out of 83 patients, 34.9% were diagnosed with intussusception.
• Key predictors included absence of fever (aOR 4.4), intermittent abdominal pain/irritability (aOR 3.4), absence of diarrhea (aOR 8.0), and RUQ tenderness (aOR 8.2).
• The model achieved 66% sensitivity, 83% specificity, 68% PPV, 81.5% NPV, and an AUROC of 0.78, with 100% sensitivity and NPV at a score of ≥1.

Conclusions:

• A prediction model based on specific symptoms effectively identifies children at risk for intussusception.
• This model supports timely ultrasound decisions in pediatric patients with ambiguous abdominal symptoms.
• Improved diagnostic accuracy can lead to better patient outcomes for intussusception.